Standards and the economy

Transcription

1 Standards and the economy Prepared for Standards Australia Centre for International Economics Canberra & Sydney July 2006

2 ABOUT THE CIE The Centre for International Economics is a private economic consultancy operating out of Canberra and Sydney. It undertakes economic analysis for clients around the world. The CIE solves problems for clients by rigorously analysing markets and regulations, appraising risks and evaluating strategies. We build economic and strategic frameworks to distil complex issues to their essentials. In this way we are able to uncover new insights about emerging developments and assess payoffs from alternative strategies. The firm has been operating since Contact details are set out below and more information on what we do and our professional staff can be obtained from our website at The CIE also co-produces a quarterly report called Economic Scenarios. This analyses global risks and scenarios and can be accessed from CANBERRA Centre for International Economics Ian Potter House, Cnr Marcus Clarke Street & Edinburgh Avenue Canberra ACT 2601 GPO Box 2203 Canberra ACT Australia 2601 Telephone Facsimile cie@thecie.com.au Website SYDNEY Centre for International Economics Suite 1, Level 16, 1 York Street Sydney NSW 2000 GPO Box 397 Sydney NSW Australia 2001 Telephone Facsimile ciesyd@thecie.com.au Website

3 iii Contents Summary v 1 Introduction 1 2 Standards and the economy 2 The broad landscape 2 What exactly is a standard? 5 The demand for standards 9 The supply of standards 11 The benefits of standards 13 The costs of standards 18 3 Macro evidence 19 Background standards and growth theory 19 The basic approach 20 Range of results for Australia 21 The standards data 21 Incorporating standards in the growth equations 24 Implications and caveats 27 4 Micro evidence 28 Sample preparation procedures in mining 29 Standards in electrical and water industries 33 The risk management standard 36 5 Conclusion 40 A Estimation details 41 References 43 Boxes, charts and tables 2.1 Standards by broad field 3

4 iv CONTENTS 2.2 Change in sales between 2001 and Broad age profile of standards Standards as prescriptive knowledge Overlapping roles of standards The demand for standards The value of setting standards Types of standards and their benefits Standards and innovation Measuring safety and risk benefits of standards Impact of R&D knowledge stock on total factor productivity The number of standards Five year moving average growth of new and modified standards Indexes of standards and productivity growth Change in standards versus change in productivity R&D and standards effect on TFP TFP and combined stock of knowledge Combined R&D and standards knowledge stock effect on TFP Bias and precision in sampling Illustration of benefits of sample standards Value of sampling standards by commodity Standards, inputs and outputs Change in GDP from water and electricity related standards Attitudes to risk 38 A.1 Regression results with standards and R&D as separate variables 41 A.2 Regression results with standards and R&D combined into a single knowledge stock 41

5 v Summary This report This report provides an analysis, and where possible quantification, of the impact of standards on the Australian economy. There are many diverse standards in use There are around standards associated with Standards Australia currently in use in Australia. These are extremely diverse, ranging from electrical and water industry technical standards, to information technology standards, to mechanical engineering standards, to health and safety standards, to management related standards. The stock of standards is constantly changing; around 80 per cent of standards currently published were introduced or modified in the past 10 years. forming part of the economy s information infrastructure Standards are part of a range of institutions that contribute to the diffusion of useful knowledge. They can be seen as an effective means of collecting, embodying and disseminating ideas about how to perform various activities. Standards can distil knowledge, provide a common language for discussion, underpin markets and help solve some externality problems. with a broad range of potential economic effects. Standards potentially have a variety of economic effects including increasing the scope for gains from trade, underpinning innovation and knowledge dissemination, reducing costs of production and increasing

6 vi productivity, helping provide safety outcomes and assisting with risk management. Macroeconomic data suggests some important relationships Examining standards at an aggregate level using a statistical approach that tries to explain changes in economywide productivity, indicates that there is a relationship between the stock of standards and productivity. If standards are specified as a separate variable, a 1 per cent increase in the stock of standards is associated with a 0.17 per cent increase in economywide productivity. If standards are specified as contributing to the stock of knowledge jointly with R&D expenditure, then a 1 per cent increase in this joint stock of knowledge leads to a 0.12 per cent increase in economywide productivity. There are many caveats to this type of analysis. It does indicate, however, that there are good arguments for viewing standards as part of the growth and dissemination of knowledge in the economy. Case studies further indicate the effects of standards Four standards or groups of standards were examined as case studies in this report including: sample standards in the mining industry; standards in water and electrical industries; and the risk management standard. Sampling standards in the mining industry increase accuracy and precision in the knowledge of the true mineral content of shipments of ores and concentrates. This type of standard enhances trade, and we conservatively estimate that this generates benefits of between $24 million and $100 million per year. The variety of standards associated with the water and electrical industries have two broad effects they increase the effectiveness of the water and electricity providers in using various inputs to their operations that is, they help in establishing the networks; and

7 vii they increase the effectiveness of the water and electricity users in accessing the networks. These effects combined are estimated to generate economywide benefits of around $1.9 billion per year. There are, of course, many caveats around these estimates, particularly the difficulty in imagining or estimating what would happen the absence of standards. There is considerable scope for more information There is considerable scope for further data collection to provide a basis for ongoing evaluation of the impact of standards on the economy.

8

9 1 1 Introduction The range of documents, processes, committees and publications that constitute the world of standards is both broad and diverse. Standards cover most areas of economic activity and come in a variety of forms with a variety of effects. Although economic impacts are not necessarily the drivers of all standards, the resources devoted to the production of standards, and the ways in which standards change the behaviour of those that use them, are subject to them, or would like to improve them, will inevitably have economic effects. Because standards whether home grown and developed by Standards Australia or international and implemented through organisations such as the ISO lie behind so many activities, separating the role of standards and measuring their impact is extremely difficult. On the one hand, the diversity of standards makes it difficult to sensibly aggregate them into a single measure; while on the other hand, the economywide rather than the specific effect of standards may be the most important. This report provides an attempt to draw out some of the key characteristics of standards and to provide a broad indication of the ways in which standards influence the economy. Chapter 2 provides an overview of the role of standards in the economy, the core ideas of which are further tested through economywide statistical analysis in chapter 3, and through the examination of a number of case studies in chapter 4. Chapter 5 provides some broad conclusions.

10 2 2 Standards and the economy The broad landscape The database on standards available from SAI Global and Standards Australia allows standards to be aggregated by broad field, as well as by the committee that is responsible for producing the standard. Chart 2.1 shows the number of Australian standards by broad field. The field definitions are those used by Standards Australia and SAI Global in compiling data on standards published and sales of standards. The largest number of standards is in the electrotechnical field. These include a wide variety of standards for electrical equipment, wiring rules and so on. Many of these standards are relatively recent, or have been recently amended, although some go back as far as Within this field, the largest number of sales of standards (and related publications, between 2001 and 2005) was associated with the committee concerned with the wiring rules. The next largest number of standards is in information technology. Again these standards are extraordinarily diverse and relate to a range of products within the IT industry. The age of these standards varies, although the majority have been published or modified within the past 10 years. Within this category, the largest number of standards (and related publications) is associated with the committee concerned with the interconnection of information systems. The mechanical engineering standards (the third most numerous) are also extraordinarily diverse ranging from screws and flanges to abrasives and gas cylinders. It also includes standards relating to silo designs and tractors. Within this category, the largest number of sales of standards (and related publications, between 2001 and 2005) are associated with committees concerned with ventilation and air conditioning and pressure equipment. Health and consumer standards are the seventh and eighth most numerous and again cover a range of areas. Health standards include standards for

11 3 2 dental and medical equipment, surgical materials, imaging, radiography and ultrasonics. Consumer standards cover helmets, filing cabinets, lawn mowers, exercise bikes and more. Quite different to these product-based standards are the management and business standards, which include standards on engaging and contracting consultants, knowledge management, quality management and risk management. Within this category, the largest number of sales is associated with risk management. 2.1 Standards by broad field Energy Conversion Systems Management Electrotechnology Radiocommunications Gas Lighting Communication Technology Welding Rubber Environment Management and Business Quality and Reliability Miscellaneous Acoustics Telecommunications Water Supply Packaging Fire Safety Textiles Metals Food Technology Plastics Timber Minerals Consumer Health Civil Engineering Building Chemical Mechanical Engineering Information Technology Electrotechnical Data source: Standards Australia

12 4 2 Chart 2.2 provides a very broad indication of the changing popularity of standards within various groups. It shows the change in the value of standards sales between 2001 and The strongest growth has been in the management and business standards, although there has also been strong growth in gas, radio communications and water supply standards. Some care should be taken in interpreting these growth figures (for example, the growth in gas standards sales is partly the result of a transfer of standards from the Australian Gas Association to Standards Australia), but they serve to confirm the diverse nature of standards and their diverse patterns of use in Australia. 2.2 Change in sales between 2001 and % 60% 40% 20% 0% -20% -40% 26% 8% 9% 2% 1% 0% -3% -3% -2% -1% -2% 0% -16% -16% Gas Acoustics Building Civil Engineering Chemical Consumer unications Technology Electrical Environment Fire Protection Food Technology Health formation Technology Lighting Data source: SAI Global Comm In 65% anagement and Business M 6% 6% 1% -8% -14% -10% -9% -12% Mechanical Engineering Minerals Miscellaneous Metals 21% -16% -11% -32% C onditions of Contract Packaging Plastics Quality and Reliability R adio Communications Rubber Occupati onal Health and Safety 9% 17% 0% -6% -4% -5% S ystems Management Specifications Telecommunicaiton s and Electrotechnology Timber Textiles Welding Water Supply Chart 2.3 illustrates the age profile of standards. It shows the proportion of the total stock of standards (at a given point in time) that were introduced or modified within the previous 5 or 10 years. The proportion introduced or modified in the previous 5 years has varied between 40 and 60 per cent, and is currently at 41 per cent. The proportion introduced or modified in the previous 10 years has varied between 75 and 96 per cent and is currently at 83 per cent. The entire stock of standards is turned over or modified roughly once every twelve years.

13 Broad age profile of standards 100% 80% 60% Share introduced or modified in previous 10 years 40% 20% Share introduced or modified in previous 5 years 0% Data source: CIE estimates What exactly is a standard? The definition of a standard used by Standards Australia is: a published document which sets out specifications and procedures designed to ensure that a material, product, method or service is fit for its purpose and consistently performs in the way it was intended. Within this definition, there is scope for a very wide range of content in the documents. Even a cursory examination of the or so standards currently in place shows that they are extraordinarily diverse. Standards say many things to many different potential users. The users too are diverse anywhere along the spectrum from consumers to producers. A more general way of characterising a standard is to say that it is a statement of how to. What this means is illustrated in chart 2.1. Using the terminology of the historian Joel Mokyr (2002), knowledge can the thought of as two types: propositional knowledge (what) and prescriptive knowledge (how).

14 Standards as prescriptive knowledge Proposition knowledge WHAT Standards (general) R&D Range of feasible techniques Prescriptive knowledge HOW to Standards (specific) R&D Costs Market pressure Actual techniques in use The full set of propositional knowledge (knowledge of the laws of physics, materials science, chemistry and so on) translates through R&D, trial and error, experiment, practice and so on into a set of feasible techniques, or knowledge how to do something. This feasible set of techniques in turn translates into the actual set of techniques in use, depending on a variety of factors such as market pressures, changing relative costs, consumer demands and so on. Standards enter into the picture by being part of the transition from the what to the how to. That is, a standard encodes within it knowledge about how to go about building things, or designing things for a purpose or how to behave in certain circumstances. For example: a product standard provides information on how to achieve a particular end with a product; a quality or design standard provides information on how to ensure that a product is adequate for a particular task;

15 7 2 a compatibility standard provides information on how to ensure that components will fit together, or that users of different information can communicate; a safety standard provides information on how to ensure that a product (or process) really does reduce risk in the area intended or how to ensure that a product is associated with minimum risk if used as intended; a process or management standard provides information on how to go about particular activities such as how to calculate how profits are tracking (earned valued standard) or how to go about managing risks (the risk management standard); and a measurement or testing standard provides information on how to go about particular measuring activities in order to ensure that the results are meaningful and can be understood by the intended users. The standard itself can convey knowledge in a prescriptive way or in a more general and performance based way. In this context, prescriptive standards are those that more tightly specify the actual technical details of the products or processes that are recommended. These types of standards convey a very specific form of how to information, which is more like a must do set of ideas. Prescriptive standards would tend to influence the actual techniques in use. Performance or more general standards provide information about what to look for when choosing a particular technique, and provide how to guidelines that can be more flexibly applied in a variety of circumstances. Distilling knowledge The process of writing a standard engaging committee members, preparing drafts, coming to consensus (in the approach that Standards Australia uses) is a process of distilling a variety of knowledge and practical experience into a document or set of documents that can be used and understood. This process is a form of information diffusion and given the resources involved developing standards, is likely to be a significant part of the many diffusion processes taking place within the economy. The case study on mineral sampling standards presented in chapter 4 is an example of how standards can be used to diffuse sound scientific knowledge into particular areas of economic activity.

16 8 2 Standards as a language Communication, particularly the communication of complex ideas, requires a common terminology, or language. A language, once established opens many possibilities for building on what is already known. Some standards particularly management and measurement related standards provide a common language to communicate key ideas. The risk management standard examined in chapter 4 is an example of a standard that provides a common language in this case a common language surrounding ideas of risk and how to manage that risk. In many markets such as risk markets the establishment of a common language can assist in allowing that market to expand and become more efficient. Other dimensions of standards Standards are not always viewed as primarily informational devices, particularly as standards are often associated with the process of standardisation and regulation. Chart 2.5 illustrates that these other dimensions of standards may overlap with the information dimension, but they do not alone capture the full picture. 2.5 Overlapping roles of standards Standardisation Information transfer Regulation

17 9 2 While standards are often associated with standardisation, it is possible to have standards without standardisation emerging: it depends entirely on the nature of the standard. To the extent that standards focus on approaches or processes (that is, to the extent to which they are performance based) they will not necessarily lead to any standardisation at all. Indeed, by providing a common basis for some activities (such as in the case of the risk management standard), standards may in fact lead to increased diversity in outcomes. Some standards, but not all, are imposed by regulation 1. When a standard is made compulsory in this sense, its nature may change again. Whereas a voluntary standard may be changed at will, once placed in regulation, a standard becomes considerably less flexible. In this case the value of the standard depends on the integrity of the regulatory process and the ability of that process to adequately weigh the benefits and costs of the regulation. Even the standardisation and regulation aspects of standards are ultimately based on the how to.. information provided by the standard itself. The demand for standards The demand for standards arises from a number of different sources as illustrated in chart 2.6. As noted, some demand for standards arises as a result of regulation. Regulation itself comes in a variety of forms and standards may be regulated for a variety of reasons. Increasingly, the most common regulation of standards arises for social reasons, in particular safety and health related reasons. In the past, much Australian regulation has tended to be anticompetitive that is providing an advantage to particular incumbents however, regulation that affects competition is currently more likely to be procompetitive. Demand for use of a particular standard (perhaps a product or design standard) may also arise as a result of market competition, or of processes that arise from various market pressures. Standards may also be a way of allowing markets to solve particular externality or public good problems. 1 Approximately one third of all Australian Standards have become mandatory under government legislation (Productivity Commission 2006, p.3).

18 The demand for standards Demand for standards Regulation Market pressures Good management Anticompetitive Other firms Shareholders Procompetitive Social Consumers Stakeholders A common example of this is the network externalities that arise when different devices (phones, faxes, DVDs etc) need to be able to communicate either with each other or with the media they are based on. In the absence of a standard, the scope for benefits from the products is considerably reduced (there is little point in 20 people having fax machines that can t communicate with each other). There is clearly a market imperative for such an externality to be resolved, as resolving it can increase the value that can be accessed by all producers. A standard can provide the means to resolve this. Whether the standard that does so must be regulated or imposed in some way is an important question, and there are mixed views on this issue. Standards can be used to help resolve underlying uncertainties within markets that may limit the scope for gains from trade. For example, the mineral sampling standards (see chapter 4), by providing more precision regarding the mineral content of ores, have the effect of increasing confidence of both buyers and sellers. Standards can also be used to underpin particular aspects of regulated markets. An efficient form of regulation for some issues may be to create a market to address the problem. For example, the greenhouse gas abatement

19 11 2 scheme established in NSW is a mandatory emissions trading scheme. The trades established in this market are, at least partly, underpinned by AS : Carbon accounting for greenhouse sinks afforestation and reforestation. Demand for standards, in particular management standards, may arise through the need for good governance and management within an organisation, or within an industry grouping. This may ultimately come from competition or through the demands of various stakeholders (shareholders, customers). The demand for standards may also arise through the demands of professionals (for example engineers, designers and so on) working in particular organisations. Particular standards may make the work of these groups more cost effective (and rewarding) leading to a demand from them for standards in some areas. The supply of standards Worldwide, there are many standard setting organisations, ranging from the well known ISO, to the less well known ITU (International Telecommunications Union) and NIST (National Institute for Standards and Technology, US). In Australia, the standards infrastructure arises through the interactions of a number of key organisations including Standards Australia, the National Associating of Testing Authorities (NATA), JAS-ANZ (the Joint Accreditation System of Australia and New Zealand), the National Measurement Institute, and various other standards writing and accreditation boards. The process Standards Australia applies to the development of new standards is as follows: Request. Standards Australia does not instigate the development of new standards but responds to requests from industry, consumers and government. Approval. Standards Australia having received a request, the need and support for a new standard is reviewed. This review assesses whether the new standard will be of benefit to the community as a whole. Technical Committee. When the new standard is approved for development a technical committee is formed. This committee is diverse in membership and consists of representatives from industry,

20 12 2 consumer groups, government, and so forth. The majority of members are unpaid and volunteer their time and resources. Aside from appointing a project manager to be the committee secretary, Standards Australia restricts its role to facilitating the development of the standard and does not actively take part in the decisions the committee makes. Draft. This is the first stage in the actual writing of the standard. At this stage the committee must ensure that the standard is not a technical barrier to trade, does not stifle innovation or competition and generally complies with other laws and regulations already in place. 2 The committee works to achieve consensus in the group and, once a draft is ready, it is made available for public scrutiny and comment. Ballot. Once all public comments have been considered and, if necessary, revisions incorporated, a ballot is held. Consensus is deemed to have been achieved if 67% of those eligible to vote have voted affirmatively and 80% of the total votes received are affirmative. There is also the requirement that no party with a major interest in the subject of the standard has voted in the negative. Two features in the writing of standards under this process are the facts that: the process draws on a wide range of expertise; and committee members volunteer their time (or are provided to the process by their employer). Chart 2.7 illustrates some of the potential tradeoffs involved in organisations volunteering resources to standards development. The value to the organisation of the new standard will depend on the value that the standard adds to the whole industry and the share that that organisation has of that industry s value. Organisations will presumably be interested in maximising both of these, although there is a potential trade-off between the effort involved in increasing value to the industry, and effort put into increasing market share. 2 It is important to note that Standards Australia does not write standards specifically for law making purposes. However it considers the use of standards in legislation and as such takes care to use terms which avoid ambiguity.

21 The value of setting standards Value to organisation = Value added to industry X Share of industry value Standards may increase total industry value Selected standards may change market share Trade-off? The benefits of standards How does the how to knowledge encoded within a standard provide economic benefits? There are a range of possibilities, depending on the type of standard and the industry or activity that it is involved with. Table 2.8 summarises some of the broad possibilities based on some of the typical classifications of standards. Particular aspects of these are examined in more detail below.

22 Types of standards and their benefits Compatibility / Interface Promotion of trade Increased consumer confidence Lowering of search costs Better knowledge dissemination Legislation Minimum Quality / Safety Promotion of trade Increased consumer confidence Lowering of search costs Better knowledge dissemination Legislation Higher market place efficiency Network externalities Variety Reduction / Focussing Devices Lowering of search costs Higher market place efficiency Information / Measurement Promotion of trade Increased consumer confidence Lowering of search costs Better knowledge dissemination Higher market place efficiency Information diffusion and innovation Research and development, an important component of any modern economy, is of little use if the ideas and discoveries that result are not diffused and used and if they can t provide a basis for future innovation. Standards, if they are timely and appropriately based, can directly contribute to the diffusion of ideas, and can provide a means by which new ideas are embodied in products or processes. Chapter 3 below considers this aspect of standards in more detail. As chart 2.9 illustrates, the effect of standards on innovation may depend on the timing of the standard. If a particular product or technical standard is imposed too early in the process of developing a new product, then the effect on innovation may be negative. On the other hand, a standard that comes along too late may result in unnecessary costs of duplication or lock in to technologies that were not the most efficient, and potentially detrimental to innovation. There is clearly a balancing act in the effect of standards on innovation, and where this balance emerges will depend on the particulars of the case. This balance will be significantly influenced by whether the standards are prescriptive or performance based. Narrow and prescriptive standards are

23 15 2 much less likely to be conducive to innovation, particularly in markets with rapid technical change. 2.9 Standards and innovation + ve Just right Effects on innovation Too early Too late ve Standards may also help innovation by providing an essential platform on which new technologies, processes and so on can build. Recent work on innovation and economic growth by the economist William Baumol (2002) suggests that the significant rate of innovation in modern economies emerges from the ability of firms to routinize innovation. Rather than random once-off innovations, modern economies are able to produce continual and routine innovations in doing things better and more efficiently 3. Routinization may itself be, at least partly, the result of firms being able to build on existing standards in products and processes. In developing, for example, a design for a new product or process, it is likely to be extremely beneficial to be able to build on standard components in developing that design. 3 This process of routine and continual development is captured by Petroski (1996) who shows the steady increments in design for some commonly used products

24 16 2 Cost of production and productivity The costs of production effects of standards are potentially significant. For processes that are routine, the use of a standard, predictable component may lower overall costs of production. As an analogy, consider the current document, which was prepared using a template that defines the way in which fonts, paragraphs, charts and tables are prepared and enter into the document. Such a template significantly lowers the cost of preparing a document, as the elements are predefined and do not need to be developed time and time again. This effect is evident in the water and electricity standards examined in chapter 4. These standards allow lower costs than would otherwise have been the case. Standards can also lower production costs by providing an efficient means of communicating particular ideas within an organisation. For example, the various management standards (quality management, risk management, earned value and so on) can be seen as an efficient primer on a particular subject and so can reduce the cost of spreading particular ideas within an organisation. Safety outcomes and risk management Many standards are designed not with specific economic outcomes in mind, but with increasing safety or managing risk. To the extent that accidents or risks lead to economic costs, these standards will have an economic impact, although this impact may be extremely hard to measure. Chart 2.10 illustrates the measurement problem. The left panel shows two alternative states of the economy with and without risk (or with and without accidents). The lines could represent the evolution of GDP or some other measure of economic activity. If the standard is successful in reducing risk, then the economy will move along the solid line, rather than the dotted line. It will have higher GDP than otherwise, but this difference will not be directly observable because the dotted line is something that didn t actually happen, and the fact that it didn t was a result of the standard 4. 4 This is not meant to imply that the standard by itself can eliminate risk, but to illustrate the point.

25 17 2 An alternative situation is illustrated in the right panel of chart The economy may be moving along a particular path, perhaps suffering from many safety issues, but after the introduction of the standard, there is a measurable increase in the growth rate of the economy 5. In this case it may be possible to measure the gap between what would have continued to happen in the absence of the standard compared with what actually happened with the standard in place. This increased growth rate may be associated with higher productivity, which is examined further in chapter Measuring safety and risk benefits of standards $ Economy without risk $ Unobservable gap Economy with risk Observable change in rate of growth Standard introduced Measurable gap Time Time Network externalities As noted above, the existence of network externalities may be a significant driver for the development of standards, and so may also be a major source of benefits. Interconnection standards of various kinds (whether physical or virtual) allow more people to join a particular network without threatening the integrity of the network, and therefore increase the value of the network to all users. 5 Actually measuring such an increase poses a significant empirical problem, but the principle remains.

26 18 2 The nature of these externalities means that they need to be measured and evaluated at an aggregate level, as the gains to any individual user will not capture the full extent of the benefit. At the economywide level, capturing network externalities may appear as an aggregate productivity increase. This is examined further in chapter 3. In some cases, such as water and electricity networks, these gains may be within the network but transmitted through the rest of the economy through costs savings. This is examined further in chapter 4. The costs of standards Standards, of course, do not come for free. Significant resources must be devoted to developing them and there may be significant tradeoffs in the ways in which standards lead to behavioural changes. While the time devoted to standards committees and so on is volunteered, it is still a cost to the economy as those individuals could have been doing something else. Further, to the extent that standards are inappropriately regulated perhaps by being too prescriptive they may in fact impose costs on the economy. A lot of Australia s microeconomic reform throughout the 1980s can in part be seen as a process of eliminating overly restrictive standards of various kinds. There may also be unexpected tradeoffs that emerge from the behavioural effects of particular standards.

27 19 3 Macro evidence Background standards and growth theory In recent analyses of the causes of economic growth, there is increasing recognition of the role of knowledge of various kinds as a major driver of productivity growth, which is in turn a major driver of economic growth. 6 Given that in many ways standards can be seen as a form of knowledge, and in particular as a way of disseminating knowledge, it is natural to consider analysing the effect of standards within the framework of growth theory. Indeed, as noted above, many of the effects of standards may only be visible at an aggregate level and may ultimately emerge as improved economywide productivity. Analysing standards within the framework of growth theory brings both benefits and significant challenges. On the benefits side, such analysis if successful would allow standards to be compared with other forms of knowledge growth (for example R&D) as well as providing an indication of the contribution that standards make to economic growth. This is the approach recently taken by DTI in their 2005 publication The Empirical Economics of Standards. There are, however, significant challenges in undertaking such analysis. First, empirical applications of growth theory are very mixed in their results. The estimated effects of various drivers depend very much on the data sets used and on the specification of the models estimated. Recent work at the Productivity Commission (Shanks and Zheng, 2006) shows how extensive this challenge can be. The major reason for this difficulty is statistical. The empirics of growth theory mean trying to discern causal relationships between variables that have a strong underlying trend. The always present risk is that various spurious correlations are likely to emerge, and while there are means of testing for these, the tests are not themselves particularly powerful. 6 A recent history of economic thinking in this area can be found in Warsh (2006).

28 20 3 MACRO EVIDENCE This problem is further compounded by the fact that it is very difficult to obtain empirical measures that correspond to the theoretical notions of growth theory. Constructing a knowledge index, for example, poses many difficulties. This problem is particularly acute in the case of standards, where the main data available on a time series basis is simply the number of standards published at a particular point in time. Nevertheless, this chapter reports the results of analysis that attempts to add standards into a typical growth theory framework. The basic approach Economic growth (the increase in GDP or GDP per person over time) depends upon both the use of various factors of production (land, labour, capital) and the efficiency with which these various factors are used. Growth will increase if we use more labour, more land and more capital. But within any economy, there is a limit to the extent to which these factors can be increased. Growth will also increase if productivity, the efficiency with which the various factors of production are used, increases. Total factor productivity (TFP, sometimes called multifactor productivity or MFP) is a measure of how much is produced for a given level of total factor use. An increase in TFP means we get more for a given set of resources. One aspect of growth theory involves considering the relative importance of the use of factors versus TFP changes in generating output increases. Studies generally find that TFP is a significant contributor to economic growth. Another aspect of growth theory is to consider the importance of various factors that determine the rate of TFP growth itself. These factors include R&D, education, trade, foreign R&D and, potentially, standards. There are a number of ways in which the underlying equations of growth theory can be expressed and then empirically estimated. In the study undertaken by DTI, labour productivity was considered to be a function of the ratio of capital to labour and the stock of standards. DTI found significant positive results from this analysis. However, adopting this approach for Australia did not yield significant or positive results. An alternative approach, reported here and used recently by the Productivity Commission, is to use TFP as the dependent variable and to

29 21 3 MACRO EVIDENCE attempt to explain it with a range of determining variables. The most common of these is R&D, or more correctly, the estimated stock of knowledge emerging from R&D spending. Range of results for Australia There is a broad range of estimates of the relationship between R&D stock and total factor productivity. Chart 3.1 illustrates the probability distribution of the estimates derived from recent work at the Productivity Commission. 3.1 Impact of R&D knowledge stock on total factor productivity a Elasticities Average = per cent confidence interval: 0.04 to a This distribution is from a meta-analysis of various equations, the coefficients of which are given equal weight. The estimating equations underlying these results are of the form ln(tfp) = a + b.ln(r&d stock) + c.ln(other variables); where TFP is total factor productivity. Data source: CIE estimates based on results in Shanks and Zheng (2006) These results show, for example, that on average, a one per cent increase in the stock of R&D was found to lead to a 0.05 per cent increase in the economywide TFP. The estimates of this elasticity ranged from 0.04 to 0.07 depending on the model specification used. As Shanks and Zheng (2006) point out, these results should be treated with some caution as there are many, many unresolved issues in estimating them. However, they do provide something of a benchmark against which to judge the results of our analysis of the effect of standards. The standards data There are two historical data series available for Australian standards:

30 22 3 MACRO EVIDENCE the annual number of new and revised standards published each year; and the total stock of current standards. The data used here are from 1962 to There are some gaps in the series in early years (particularly for the stock of standards), but it is possible to fill these gaps using some basic stock-flow accounting 7. Chart 3.2 shows the cumulative stock of standards and the number of standards published in each year. Both of these have increased steadily, although there are some notable periods of acceleration and deceleration in the publication of new and revised standards. 3.2 The number of standards Cumulative standards (left axis) Data source: CIE estimates based on data from Standards Australia and SAI Global New and revised standards (right axis) Chart 3.3 shows the five year moving average of the change in the number of new or revised standards, and indicates that the late 1960s and early 1970s were times of high growth, as were the mid to late 1990s For example, the current stock of standards is equal to last year s stock plus additions and minus removals. With some very mild assumptions we can use this relationship to fill in the gaps in the series.

31 23 3 MACRO EVIDENCE 3.3 Five year moving average growth of new and modified standards 20% 10% 0% -10% Data source: CIE estimates Standards and productivity Chart 3.4 shows an index of the stock of standards along with an index of total factor productivity for the economy. Both are clearly increasing, although the stock of standards is growing considerably faster than total factor productivity. 3.4 Indexes of standards and productivity growth Index of cumulative standards Growth 4.6 per cent a year 2 Index of total factor productivity Growth 1.1 per cent a year Data source: ABS, CIE estimates based on data from Standards Australia and SAI Global Chart 3.5 shows the relationship between the annual change in standards and the annual change in total factor productivity. There is no direct

32 24 3 MACRO EVIDENCE relationship here, and statistical tests suggest that it is not sensible to relate these two variables alone. 3.5 Change in standards versus change in productivity Change in stock of standards (%) Change in total factor productivity (%) Data source: CIE estimates Incorporating standards in the growth equations There are two ways of incorporating standards into the typical growth equations. One is to treat standards as a separate stock of knowledge, and to incorporate this as a separate variable in the growth equations, alongside the stock of knowledge defined by R&D. The second approach is to treat standards as a variable which helps define the stock of knowledge in the economy and interacts with other variables such as R&D. Under this approach, a new stock of knowledge determined by R&D growth and standards growth together is used in the estimating equations. Putting standards alongside R&D Chart 3.6 summarises the results of a number of estimating equations that relate TFP to the change in the stock of R&D, and various definitions of the stock of standards. The results indicate that: a one percent increase in the stock of R&D leads to, on average, a per cent increase in TFP. This is very similar to the Shanks and Zheng (2006) results; and a one per cent increase in the stock of standards leads to a 0.17 per cent increase in the level of TFP. This estimate ranges from 0.14 to 0.19,

33 25 3 MACRO EVIDENCE which does not intersect with the range for the R&D estimate, indicating that they are significantly different. 3.6 R&D and standards effect on TFP: Range of elasticity estimates a R&D effect Average = per cent confidence interval: 0.04 to 0.07 Standards effect Average = per cent confidence interval: 0.14 to a Summary based on four estimating equations: ln(tfp) = a + b.ln(r&d Stock) + c. ln(standards stock). The equations have slight variations in the definition of the stock of standards and in the inclusion of foreign R&D variables. The R 2 for the estimating equations are 0.97, t statistics for the R&D variable range from 3 to 5.5 and t statistics for the standards variable range from 4.8 to The variables appear to be cointegrated. See appendix for more details. Data source: CIE estimates The statistical performance of these estimates appears acceptable, at least as acceptable as any other estimations of this kind. Nevertheless, there is still a risk that these estimates are spurious, which should be kept in mind in the interpretation discussion below. Combining standards and R&D into a single knowledge stock The second approach is to experiment with combining both R&D expenditure and the development of new or modified standards to form a single combined index of the stock of knowledge. Chart 3.7 shows the relationship between this combined stock of knowledge and the index of total factor productivity (both expressed in logs). The chart indicates that there is a strong positive relationship.

34 26 3 MACRO EVIDENCE 3.7 TFP and combined stock of knowledge Combined stock of knowledge (log) Total factor productivity (log) Data source: CIE estimates Chart 3.8 summarises the results of estimating equations incorporating the combined stock of knowledge variable. The results show that, on average, a one per cent increase in the combined stock of knowledge (standards and R&D) leads to a 0.12 per cent increase in total factor productivity. The 95 per cent confidence interval is from 0.10 to 0.13, which indicates that this elasticity is fairly precisely estimated. 3.8 Combined R&D and standards knowledge stock effect on TFP a Average = per cent confidence interval: 0.10 to a Summary based on six estimating equations ln(tpf) = a + b.ln(combined stock). There are differences in the definition of the combined stock and the inclusion of foreign R&D between the equations. The R 2 for the equations are around The t statistic for the stock variable ranges from 5.18 to The variables appear to be cointegrated. See appendix for details. Data source: CIE estimates

35 27 3 MACRO EVIDENCE Implications and caveats Several points emerge from the results presented here. There is apparently a positive relationship between economywide total factor productivity and the stock of standards, either when this is kept as a separate variable, or combined with a stock of R&D variable. This relationship is of an economically significant magnitude. These results are consistent with the idea that standards act as a means of diffusing or embodying knowledge. The various estimating equations perform relatively well at a statistical level, certainly as well as other similar equations in the growth literature. In the detailed estimates (reported in the appendix), there is an interesting interaction between the stock of standards and the stock of foreign R&D. These variables appear to play a similar role, or to be related to a similar but unobserved third variable. At the same time these results should be treated with a great deal of caution. It is difficult to be completely confident in the statistical properties of the estimates generated here. It remains possible that the estimates are not in fact meaningful although this conclusion would need to be applied to all estimates of this kind. Importantly, the standards variable used here is simply the cumulative stock of standards, or in some cases the total number of new or modified standards published in a given year. While this broad variable should be related to the knowledge content of standards, there is no guarantee that this is so. Even if it is, it is a crude measure as there are clear differences between particular individual standards and potentially large differences in their significance.

36 28 4 Micro evidence This chapter presents case studies of four groups of standards in order to further illustrate the ways in which standards have economic effects and where possible to quantify the magnitude of the effects. Case studies provide an alternative to the broad macroeconomic analysis presented in the previous chapter as a way of examining the impact of standards. The challenge with case studies, however, is that because of the diversity of standards, it is very difficult to construct case studies that are representative of all standards. The case studies presented here were chosen not in order to be representative, but in order to illustrate particular aspects of the economic effects of standards. These aspects can be summarised as follows. The standards around sample preparation procedures in mining illustrate the ways in which standards can be used to disseminate sound scientific knowledge and principles. These standards also illustrate the role that Australia can play in international standard setting in a way that increases the scope for gains from trade. The various standards in the water and electrical industries illustrate the ways in which standards can enhance the productivity of particular activities. Given the importance of both these industries in the economy, these cases also illustrate the ways in which the effects of standards can be transmitted throughout the economy. The risk management standard, one of the best selling Australian Standards, also illustrates the way in which standards can be used to collate and transmit valuable information. At the level of individual standards, or groups of standards, a significant difficulty in estimating economic effects is in constructing estimates of what would happen without the standard. This counterfactual can never be directly observed, and must be imputed often subjectively through general knowledge of the operations of a particular industry.